"These findings illustrate the need for assessing p-tau, in addition to Aβ, for clinical trial recruitment and design," first author Rahul Desikan told Alzforum in an e-mail.

Desikan and colleagues used data from 107 clinically normal patients enrolled in the Alzheimer's Disease Neuroimaging Initiative (ADNI). Each had CSF samples taken at baseline. They also underwent cognitive analysis then, and three years later. Cognitive tests included the Clinical Dementia Rating Scale (CDR), CDR-Sum of Boxes (CDR-SB), and the Alzheimer's Disease Assessment Scale-Cognitive subscale (ADAS-cog). Based on CSF analysis, the research group categorized subjects as having high (>23 pg/mL) or low (<23pg/mL) p-tau levels, and high (>192 pg/mL) or low (<192 pg/mL) Aβ42. Those cutoff values were previously reported to predict progression from mild cognitive impairment (MCI) to AD (see Shaw et al., 2009).

By itself, a low CSF Aβ42 level at baseline did predict, on average, cognitive decline in these normal individuals, with all three cognitive test scores dropping. But the scientists picked that relationship apart by examining whether CSF p-tau influenced the interaction. They found that cognitive scores actually fell only in those people who had both low CSF Aβ and high p-tau. If CSF p-tau levels were low, regardless of Aβ level, scores were closer to normal. "If you take p-tau away, there's no effect of Aβ on clinical decline," said Desikan. "This suggests that p-tau is acting like a critical link between Aβ and clinical decline." Similar findings were reported previously (see Fagan et al., 2009).

The findings should help researchers select patients for prevention trials. One of the problems researchers face when conducting those trials is finding at-risk individuals who will progress to dementia before the trial is over. Screening for elevated CSF p-tau levels in addition to lowered Aβ42 may help scientists catch cognitively normal people in the two to five years before they begin to decline, wrote David Holtzman, Washington University, St. Louis, Missouri, in an accompanying editorial. Using such a sample population would allow trials (which normally last only 18 months) a better chance of observing a drug's preventative capabilities.

On the other hand, researchers may prefer to select people who have low CSF Aβ but not yet have elevated tau levels, Desikan said. If tau phosphorylation is initiated by Aβ but then continues independently of Aβ (dependence on Aβ is not yet known), then that would imply that it is too late for Aβ-clearing drugs to stop the neurodegenerative process. The bottom line, he said, is that, for any given study, researchers should know the p-tau status of their participants.

Results were more ambiguous for high total tau (t-tau), which did not link Aβ and cognitive decline, according to CDR and CDR-SB results, though it did in the ADAS-cog. In previous studies, t-tau has been shown to correlate well with p-tau (see Fagan et al., 2011). Differences in CSF assays could cause the discrepancy between this study and previous ones, or p-tau might be a stronger biomarker of AD than t-tau, Desikan said.

Similar to the current findings, Desikan and colleagues last year reported that brain atrophy only occurs in non-demented healthy controls or people with MCI who have both low CSF Aβ42 and high p-tau, not just low CSF Aβ42 (see ARF related news story on Desikan et al., 2011). Other studies hinted at biomarkers other than CSF Aβ that predict cognitive decline in healthy controls such as the p-tau or t-tau to Aβ42 ratio (see Fagan et al., 2007), YKL40 (chitinase-3 like-1; see Craig-Schapiro et al., 2010), or visinin-like protein 1 (see Tarawneh et al., 2011). Taken together, recent evidence suggests that CSF biomarkers other than just Aβ42 will be important for screening in clinical trials.

"There are some in the field who seem to want to select patients based on either CSF amyloid markers or PET scan evidence of amyloid deposition for inclusion in 'prevention' trials," wrote Peter Davies, Albert Einstein College of Medicine in the Bronx, New York, in an e-mail to Alzforum. "This paper and the majority experience of autopsy studies suggest that such a group would have a low likelihood of progression, and that inclusion of p-tau markers would really help."

CSF p-tau, along with Aβ42, probably won't be used in recruitment of secondary prevention trials right away, said Eric Reiman, Banner Alzheimer's Institute, Phoenix, Arizona, because he and other researchers are initially looking for a range of pathology in cognitively normal people, including very early amyloid deposition without neurodegeneration. The presence of p-tau indicates that neurons have already begun to die, he said, and some treatments could be too little too late at that point. But depending on the type of participants needed, later study recruitment will likely use such biomarkers of neurodegeneration, said Reiman.—Gwyneth Dickey Zakaib.

References:
Desikan RS, McEvoy LK, Thompson WK, Holland D, Brewer JB, Aisen PS, Sperling RA, Dale AM. Amyloid-β-associated clinical decline occurs only in the presence of elevated p-tau. Arch Neurol 2012 April 23. Abstract

Holtzman D. CSF Biomarkers for Secondary Prevention Trials: Why Markers of Amyloid Deposition and Neurodegeneration Are Both Important. Arch Neurol 2012 April 23. Abstract